Export the Data in a Format Appropriate for Your Analyses¶
From the beginning of the coding process, keep in mind the data you want to analyze and how you want your data formatted for analyses. Your coded data are what you will analyze. So, you should be sure that the way you code your data is compatible with the way that you will analyze it. Think about how you want your data to look when you export it from Datavyu to analyze it elsewhere (e.g., Excel, SPSS).
Although we recommend using alphabet letters as codes rather than numbers, most researchers prefer to analyze numeric data rather than strings. Although onset and offset times contain the information you need to understand timing relations, most researchers prefer to analyze durations rather than relative times.
Datavyu’s Export File function will export the strings and raw onset/offset times from the spreadsheet. You can convert these data into numeric and durations in your analysis spreadsheet (with simple compute functions in Excel or SPSS, for example). But you can also export the data in these formats using Datavyu scripts.
Most researchers like to analyze their data in square spreadsheet formats. Many researchers organize the data with one row for the smallest unit of analysis (e.g., a trial, a look, a touch, a facial gesture) and many rows for each participant. Some researchers maintain one row per participant and organize tasks or trials across columns. Some researchers aggregate the data (e.g., by averaging over trials) prior to export to maintain one row per participant.
Behavioral data have a naturally nested structure: Trials are nested within conditions; conditions in turn are nested within sessions and participants. Facial expressions are nested within interactions and interactions are nested in turn within particular situations within the session. Behavioral data are naturally interleaved: One event is ongoing while another event is starting or ending.
Datavyu’s temporal alignment feature provides coders with immediate information about the nested and interleaved temporal structure of events. Thus, we recommend that coders delineate the structure of the session by coding cells to represent the larger and smaller nested units. However, the nested and interleaved structure of events cannot be exported automatically without the user specifying with a script how they would like to see the events formatted. Although Datavyu has an automatic Export File function, it will not repeat information down rows of data unless the data are arranged like that in the spreadsheet. There is no need to repeat participant ID or condition labels across every trial, however, because you can request this with a script for exporting the data. Moreover, Datavyu’s automatic Export File function will not carve overlapping and interleaved events into mutually exclusive categorical combinations (e.g., of talking, touching, and looking). You will need to do this with an export script.
This video displays a user running a script to export data in a specific way. It exports all of the columns of one spreadsheet into an excel file. Instead of just exporting one cell from a Datavyu spreadsheet, into one cell in excel, it makes it possible to repeat important information that you want to store in multiple cells. In this example, participant metadata (id, birthdate, testdate) is stored in only one cell in a Datavyu spreadsheet but it is information that gets repeatd down multiple rows of data to make it potentially easier to analyze in a statistical program.